Our study reports findings from all registered suicides among the total Norwegian prison population within a 17-year period. During 2000–2017, suicide was the leading cause of death during imprisonment and among the leading causes of death after release. The mean suicide rate was 133.8 per 100 000 in prison, and 82.8 per 100 000 after release. In comparison, the annual global age-standardized suicide rate of 11.4 per 100 000 population in 2012 .
The immediate periods after imprisonment and after release both imposed a high risk of suicide; the suicide rate on day 1 was almost seven times higher than week 1, and about ten times higher than months 2–6. Although the suicide rates after release were lower, we also found a similar trend related to time after release; the risk of suicide on day 1 after release was more than twice as high as week 1 after release, and six times higher than 2–6 months after release.
Our results documenting a peak in suicides immediately after imprisonment are in line with prior international findings showing an increased risk during the first weeks and days in prison [18, 33]. We also found an increased suicide risk among people on pre-trail detention and in high-security units supporting the existing international literature [34–37].
Although the suicide rates were lower after release, we found a doubled risk of suicide on day 1 post-release, compared to the rest of the week. With the exception of one study taking place in the northeast of Australia , our findings are in line with previous studies finding a peak in suicides after release [7, 18]. The high overall suicide rate supports the idea that the transition to life outside prison is a period with substantially increased risk for premature death. Previous research has also suggested increased risks for overdose death during the immediate period post release [4, 22].
One main finding was the association between suicide both in prison and after release and being convicted of murder: the risk of suicide in prison was more than 27 times higher among people convicted of murder, adjusted for other factors. The link between violent crimes and suicide has also reported in other studies [13, 15, 16]. Radeloff and colleagues found the highest risks for suicide among offenders convicted for offences against life, bodily integrity or against sexual self-determination .
In a recent meta-analysis of prisoner suicide rates in 24 high-income countries in Europe, Australasia, and North America, Fazel and colleagues found that the rates of prisoner suicide were higher in countries where fewer individuals were imprisoned per 100 000 members of the general population . The authors proposed an explanation for this link related to the prisoners being more selected in terms of having sentences for more serious or violent offences and more likely to be suffering from mental illnesses, and thus more vulnerable.
Strengths and Limitations
Using mandatory national registries is a major strength of the study. The datasets are linked using unique 11-digit identifiers assigned to all residents in Norway, minimizing the risk of linkage-biases. Moreover, all deaths are classified according to the most recent ICD criteria, and death-categories are reported according to individual ICD codes, minimizing the risk of information bias.
Having a national cohort followed for 17 years enables stratified analysis, which is another major strength of the study. Our study is based on a large sample, and our results advance more precise day-by‐day understanding of risk of suicide following incarceration and release. However, suicide is a rare event, and when analysing stratified groups, such as suicides per main conviction, some groups will have small numbers. This results in higher uncertainty, reflected in wide confidence intervals.
Another limitation is the lack of demographic and socio-cultural variables in our dataset, in addition to information on mental health – factors that are associated with suicide in prisoners .
Misclassifications of causes of death may occur in registry data: it may be that some suicides might be classified as overdose deaths or accidents, causing an underestimation. However, Norwegian data has been assessed as having good validity and reliability for suicide classification .
Moreover, the number of suicides in prison may be somewhat underestimated due to how suicides are recorded in the databases. A person may have initiated a suicidal act in prison but dies in hospital later. Such mortalities will not be recorded as in-prison suicides in this article, as death occurred outside of prison.